introduction to qcd on the lattice
TRANSCRIPT
Why Basics Chall EFTs Rev Future
Introduction to QCD on the lattice
Rainer Sommer
DESY, Zeuthen
September 2005 (DESY, Hamburg)
Why lattice QCD
Basic steps
Challenges
Effective Field Theories
A revolution?
The future
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
An Introduction?
I there are about 1.5m of proceedings of annual lattice conferencessince ’85(?)
I there are ≥ 4 text-books
I so what can I say?
if it is not what you need, please ask ...
... or open the Monvay/Munster
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
An Introduction?
I there are about 1.5m of proceedings of annual lattice conferencessince ’85(?)
I there are ≥ 4 text-books
I so what can I say?
if it is not what you need, please ask ...
... or open the Monvay/Munster
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
An Introduction?
I there are about 1.5m of proceedings of annual lattice conferencessince ’85(?)
I there are ≥ 4 text-books
I so what can I say?
if it is not what you need, please ask ...
... or open the Monvay/Munster
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
An Introduction?
I there are about 1.5m of proceedings of annual lattice conferencessince ’85(?)
I there are ≥ 4 text-books
I so what can I say?
if it is not what you need, please ask ...
... or open the Monvay/Munster
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
An Introduction?
I there are about 1.5m of proceedings of annual lattice conferencessince ’85(?)
I there are ≥ 4 text-books
I so what can I say?
if it is not what you need, please ask ...
... or open the Monvay/Munster
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
Why do we work on lattice QCD
it provides
I a rigorous definition of QCD
I a computational tool for QCD in the non-perturbative regime−→ numbers for phenomenology
I the connection between perturbative and non-perturbativephenomena
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
Why do we work on lattice QCD
it provides
I a rigorous definition of QCD
I a computational tool for QCD in the non-perturbative regime−→ numbers for phenomenology
I the connection between perturbative and non-perturbativephenomena
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
Why do we work on lattice QCD
it provides
I a rigorous definition of QCD
I a computational tool for QCD in the non-perturbative regime−→ numbers for phenomenology
I the connection between perturbative and non-perturbativephenomena
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A rigorous definition of QCD
Wilson regularization
I strict positivity [Luscher;Creutz ]renormalizability (≡ contin. limit) shown to all order of PT [Reisz ]all (flavor) vector symmetries
I axial symmetries have to be restored by proper renormalization [Bochicchio et al. ](as in dim regularization)
Ginsparg Wilson regularization
I exact chiral symmetry in the regularized theory(no questions about γ5)
I can prove– existence of topological suszeptibility [Giusti,Rossi,Testa; Luscher ]– Witten-Veneziano formula– index theorem
I Conceptually clean framework if there ever is a doubt about an n-loopcomputation
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
A computational tool from “first principles”
I first principles
LQCD = − 1
2g20
tr FµνFµν+∑
f
ψf D + mf ψf
Experiment︷ ︸︸ ︷Fπmπ
mK
mD
mB
LQCD(g0,mf )=⇒
QCD parameters (RGI)︷ ︸︸ ︷ΛQCD
M = (Mu + Md)/2Ms
Mc
Mb
+
Predictions︷ ︸︸ ︷ξ
FB
BB
...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
What doesLQCD(g0,mf )
=⇒ mean?
Discretization of LQCD
with
I gauge invariance
I locality
I unitarity
=⇒
renormalization ⇓ continuumlimitlow energy matrix elements
±O
„1
√computer time
«
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
What doesLQCD(g0,mf )
=⇒ mean?
Discretization of LQCD
with
I gauge invariance
I locality
I unitarity
=⇒
renormalization ⇓ continuumlimitlow energy matrix elements
±O
„1
√computer time
«
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
What doesLQCD(g0,mf )
=⇒ mean?
Discretization of LQCD
with
I gauge invariance
I locality
I unitarity
=⇒
renormalization ⇓ continuumlimitlow energy matrix elements
±O
„1
√computer time
«
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Why Def tool
What doesLQCD(g0,mf )
=⇒ mean?
Discretization of LQCD
with
I gauge invariance
I locality
I unitarity
=⇒
renormalization ⇓ continuumlimitlow energy matrix elements
±O
„1
√computer time
«
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Basic steps of a computation
this was a bit simplified ... we need the basic steps
I Formulate the problem in the Euclidean
I Discretize ≡ Regularize and simulate
I Renormalize
I Continumm limit (remove the Regularization)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Basic steps of a computation
this was a bit simplified ... we need the basic steps
I Formulate the problem in the Euclidean
I Discretize ≡ Regularize and simulate
I Renormalize
I Continumm limit (remove the Regularization)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Basic steps of a computation
this was a bit simplified ... we need the basic steps
I Formulate the problem in the Euclidean
I Discretize ≡ Regularize and simulate
I Renormalize
I Continumm limit (remove the Regularization)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Basic steps of a computation
this was a bit simplified ... we need the basic steps
I Formulate the problem in the Euclidean
I Discretize ≡ Regularize and simulate
I Renormalize
I Continumm limit (remove the Regularization)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Easy (for the formulation!) case:
0A A0
Spectrum and simple matrix elements
an example: flavour currents: −→ time
Aijµ = ψiγµγ5ψj , ipµFK = 〈K (p)|Aus
µ (0)|0〉
Z 2A
∫d3x 〈Aus
0 (x) Asu0 (0)〉 = − 1
2FK2mKe−x0mK
+O(e−(m′K−mK)x0) + O(e−Lmπ )
another example5γ5 γx
0
y
s = 2∆
∫d3xd3y 〈Asu
0 (x)O∆s=2(0)Asu0 (y)〉
d3x 〈Aus0 (x)Asu
0 (0)〉∝ BK
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Easy (for the formulation!) case:
0A A0
Spectrum and simple matrix elementsan example: flavour currents:
−→ time
Aijµ = ψiγµγ5ψj , ipµFK = 〈K (p)|Aus
µ (0)|0〉
Z 2A
∫d3x 〈Aus
0 (x) Asu0 (0)〉 = − 1
2FK2mKe−x0mK
+O(e−(m′K−mK)x0) + O(e−Lmπ )
another example5γ5 γx
0
y
s = 2∆
∫d3xd3y 〈Asu
0 (x)O∆s=2(0)Asu0 (y)〉
d3x 〈Aus0 (x)Asu
0 (0)〉∝ BK
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Easy (for the formulation!) case:0A A0
Spectrum and simple matrix elementsan example: flavour currents: −→ time
Aijµ = ψiγµγ5ψj , ipµFK = 〈K (p)|Aus
µ (0)|0〉
Z 2A
∫d3x 〈Aus
0 (x) Asu0 (0)〉 = − 1
2FK2mKe−x0mK
+O(e−(m′K−mK)x0) + O(e−Lmπ )
another example5γ5 γx
0
y
s = 2∆
∫d3xd3y 〈Asu
0 (x)O∆s=2(0)Asu0 (y)〉
d3x 〈Aus0 (x)Asu
0 (0)〉∝ BK
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Easy (for the formulation!) case:0A A0
Spectrum and simple matrix elementsan example: flavour currents: −→ time
Aijµ = ψiγµγ5ψj , ipµFK = 〈K (p)|Aus
µ (0)|0〉
Z 2A
∫d3x 〈Aus
0 (x) Asu0 (0)〉 = − 1
2FK2mKe−x0mK
+O(e−(m′K−mK)x0) + O(e−Lmπ )
another example5γ5 γx
0
y
s = 2∆
∫d3xd3y 〈Asu
0 (x)O∆s=2(0)Asu0 (y)〉
d3x 〈Aus0 (x)Asu
0 (0)〉∝ BK
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I easy case: Equilibrium thermodynamicsfinite temperatureEuclidean L0 × L3 spacetime,
ψ(x+L00) = −ψ(x) , ψ(x+L00) = −ψ(x)
Aaµ(x+L00) = +Aa
µ(x)
then path integral = Tre−HL0 = Tre−H/T
L0 = 1/T , T = Temperature
and finite chemical potential µq
LQCD → LQCD + µq ψγ0ψ︸ ︷︷ ︸number density
path integral = Tre−(H−µqNq)/T
−→ thermodynamic properties, eg. EOS talk by S. Katz
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I easy case: Equilibrium thermodynamicsfinite temperatureEuclidean L0 × L3 spacetime,
ψ(x+L00) = −ψ(x) , ψ(x+L00) = −ψ(x)
Aaµ(x+L00) = +Aa
µ(x)
then path integral = Tre−HL0 = Tre−H/T
L0 = 1/T , T = Temperature
and finite chemical potential µq
LQCD → LQCD + µq ψγ0ψ︸ ︷︷ ︸number density
path integral = Tre−(H−µqNq)/T
−→ thermodynamic properties, eg. EOS talk by S. Katz
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Difficult case: particle decays, scatteringrelation phase shifts ⇔ spectrum in a finite L3 box
[Luscher; Luscher&Wolff; ... ]
I Difficult case: Matrix elements with more than one hadrons ininitial/final statenogo [Maiani & Testa ]go [Lellouche & Luscher ] Matrix elements in a finite L3 box
talks by Giusti, Lubicz
I Impossible (?) casesH H’ → 5 HThe mass of a di-quarkThe gluon condensate...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Difficult case: particle decays, scatteringrelation phase shifts ⇔ spectrum in a finite L3 box
[Luscher; Luscher&Wolff; ... ]
I Difficult case: Matrix elements with more than one hadrons ininitial/final statenogo [Maiani & Testa ]go [Lellouche & Luscher ] Matrix elements in a finite L3 box
talks by Giusti, Lubicz
I Impossible (?) casesH H’ → 5 HThe mass of a di-quarkThe gluon condensate...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Difficult case: particle decays, scatteringrelation phase shifts ⇔ spectrum in a finite L3 box
[Luscher; Luscher&Wolff; ... ]
I Difficult case: Matrix elements with more than one hadrons ininitial/final statenogo [Maiani & Testa ]go [Lellouche & Luscher ] Matrix elements in a finite L3 box
talks by Giusti, Lubicz
I Impossible (?) casesH H’ → 5 HThe mass of a di-quarkThe gluon condensate...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Formulation
I Difficult case: particle decays, scatteringrelation phase shifts ⇔ spectrum in a finite L3 box
[Luscher; Luscher&Wolff; ... ]
I Difficult case: Matrix elements with more than one hadrons ininitial/final statenogo [Maiani & Testa ]go [Lellouche & Luscher ] Matrix elements in a finite L3 box
talks by Giusti, Lubicz
I Impossible (?) casesH H’ → 5 HThe mass of a di-quarkThe gluon condensate...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Discretize and “simulate”lattice: xµ = anµ, nµ ∈ ZZ
quarks: ψ(x) on lattice points
gluons: U(x , µ) = P expn
aR 10 dtAµ(x + a(1− t)µ)
o∈ SU(3) on linksEuclidean action: S = SG + SF
SG =1
g20
Xp
tr 1− U(p) ,
SF = a4X
x
ψ(x)(D(U) + m)ψ(x)
D(U) : discretized Dirac operator
Observables:
〈O〉 = 1Z
ZD[ψ]D[ψ ]D[U] O e−S
Z : 〈1〉 = 1
Can be interpreted as a statistical system in 3 + 1 = 4 dimensions.(S → H/(kT ) , 1/(kT ) ∝ 1/g2
0 )
First principles evaluation of 〈O〉 by MC methodstatistical errors ∝ 1/
√computer time.
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Renormalization and continuum limit: remove theregularization
For simplicity discussion for Nf = 2 mass-degenerate quarks
I In QCD: critical line with (almost) massless quarks,massless π’s;
Continuum limit: (amhadron → 0) at g0 → 0
Along RGT ≡ fixed physical quark mass:
a mhadron ∼ Chadron e−1/(2b0g20 )(2b0g
20 )−b1/2b2
0 × [1 + O(g20 )]| z
badly convergent
b0 = (11− 23Nf)/(4π)2
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I More precisely
fix m2π/m
2nucleon = 137/938← experiment
fix mnucleon = 938 MeV← experiment
⇒ a =a mnucleon
938MeV← “setting the scale”
other hadron masses and e.g. Fπ are predictions up to O(a) or O(a2) errors
bare parametersg0,m0
→ renormalized onesmπ ,mnucleon
← hadronicrenormalization scheme
I Equivalently:
fix m2π/m
2nucleon, then
mhadron
mnucleon
˛lattice
=mhadron
mnucleon
˛QCD
× [1 + O(an)]
I often: “know” that one is close to the continuum limit
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I More precisely
fix m2π/m
2nucleon = 137/938← experiment
fix mnucleon = 938 MeV← experiment
⇒ a =a mnucleon
938MeV← “setting the scale”
other hadron masses and e.g. Fπ are predictions up to O(a) or O(a2) errors
bare parametersg0,m0
→ renormalized onesmπ ,mnucleon
← hadronicrenormalization scheme
I Equivalently:
fix m2π/m
2nucleon, then
mhadron
mnucleon
˛lattice
=mhadron
mnucleon
˛QCD
× [1 + O(an)]
I often: “know” that one is close to the continuum limit
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I More precisely
fix m2π/m
2nucleon = 137/938← experiment
fix mnucleon = 938 MeV← experiment
⇒ a =a mnucleon
938MeV← “setting the scale”
other hadron masses and e.g. Fπ are predictions up to O(a) or O(a2) errors
bare parametersg0,m0
→ renormalized onesmπ ,mnucleon
← hadronicrenormalization scheme
I Equivalently:
fix m2π/m
2nucleon, then
mhadron
mnucleon
˛lattice
=mhadron
mnucleon
˛QCD
× [1 + O(an)]
I often: “know” that one is close to the continuum limit
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I More precisely
fix m2π/m
2nucleon = 137/938← experiment
fix mnucleon = 938 MeV← experiment
⇒ a =a mnucleon
938MeV← “setting the scale”
other hadron masses and e.g. Fπ are predictions up to O(a) or O(a2) errors
bare parametersg0,m0
→ renormalized onesmπ ,mnucleon
← hadronicrenormalization scheme
I Equivalently:
fix m2π/m
2nucleon, then
mhadron
mnucleon
˛lattice
=mhadron
mnucleon
˛QCD
× [1 + O(an)]
I often: “know” that one is close to the continuum limit
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I often: “know” that one is close to the continuum limit
I better: see
expl. from [Garden et al. 1999 ] quenched
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
I often: “know” that one is close to the continuum limit
I better: see
expl. from [Garden et al. 1999 ] quenched
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Reg Ren Cont
Continuum limit continued
I bewareof phase transitions
example fromtmQCD[Sharpe et al.; Munster;
Scorzato ][Farchioni et al., χLF Coll. ]
β
µ
κ = (2µκ)-1
β1 1/g02
1/m0
cont.limit
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The challenges for lattice QCD as a computational tool
large L, small a, mquark of different scales
I numerical cost of a computation
I renormalization and mixing
I heavy quarks → Lubicz
furthermore:
I short physical distances to connect to PT → Knechtli
I complex actions (finite µq) very difficult for MC-technique → Katz(Seff(U) is not real for finite µq)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The challenges for lattice QCD as a computational tool
large L, small a, mquark of different scales
I numerical cost of a computation
I renormalization and mixing
I heavy quarks → Lubicz
furthermore:
I short physical distances to connect to PT → Knechtli
I complex actions (finite µq) very difficult for MC-technique → Katz(Seff(U) is not real for finite µq)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The challenges for lattice QCD as a computational tool
large L, small a, mquark of different scales
I numerical cost of a computation
I renormalization and mixing
I heavy quarks → Lubicz
furthermore:
I short physical distances to connect to PT → Knechtli
I complex actions (finite µq) very difficult for MC-technique → Katz(Seff(U) is not real for finite µq)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The challenges for lattice QCD as a computational tool
large L, small a, mquark of different scales
I numerical cost of a computation
I renormalization and mixing
I heavy quarks → Lubicz
furthermore:
I short physical distances to connect to PT → Knechtli
I complex actions (finite µq) very difficult for MC-technique → Katz(Seff(U) is not real for finite µq)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The challenges for lattice QCD as a computational tool
large L, small a, mquark of different scales
I numerical cost of a computation
I renormalization and mixing
I heavy quarks → Lubicz
furthermore:
I short physical distances to connect to PT → Knechtli
I complex actions (finite µq) very difficult for MC-technique → Katz(Seff(U) is not real for finite µq)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The work horse of LQCD: the HMC [Duane, Kennedy, Pendleton, Roweth ]
〈O(Φ)〉 = 1Z
ZΦ
Oe−S(Φ) Φ = U(x , µ), ψ(x), ψ(x)
= 1N
NXi=1
O(Φi ) for P(Φ) ∝ e−S(Φ) > 0
local bosonic actions: Φi with Metropolis, heatbath, OR algo’s: fast
QCD: integrate out fermions (Grassmann)
〈O(U)〉 = 1Z
ZU
[O(U)]F exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local
)
rewrite with bosonic field φ
〈O(U)〉 = 1Z
ZU,φ,φ†
[O(U)]F exp(−SG(U)− Nf2φ†[(D(U) + m0)
†(D(U) + m0)]−1φ)
≡ 1Z
ZU,φ,φ†
[O(U)]Fe−SPF
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The work horse of LQCD: the HMC [Duane, Kennedy, Pendleton, Roweth ]
〈O(Φ)〉 = 1Z
ZΦ
Oe−S(Φ) Φ = U(x , µ), ψ(x), ψ(x)
= 1N
NXi=1
O(Φi ) for P(Φ) ∝ e−S(Φ) > 0
local bosonic actions: Φi with Metropolis, heatbath, OR algo’s: fast
QCD: integrate out fermions (Grassmann)
〈O(U)〉 = 1Z
ZU
[O(U)]F exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local
)
rewrite with bosonic field φ
〈O(U)〉 = 1Z
ZU,φ,φ†
[O(U)]F exp(−SG(U)− Nf2φ†[(D(U) + m0)
†(D(U) + m0)]−1φ)
≡ 1Z
ZU,φ,φ†
[O(U)]Fe−SPF
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The work horse of LQCD: the HMC [Duane, Kennedy, Pendleton, Roweth ]
〈O(Φ)〉 = 1Z
ZΦ
Oe−S(Φ) Φ = U(x , µ), ψ(x), ψ(x)
= 1N
NXi=1
O(Φi ) for P(Φ) ∝ e−S(Φ) > 0
local bosonic actions: Φi with Metropolis, heatbath, OR algo’s: fast
QCD: integrate out fermions (Grassmann)
〈O(U)〉 = 1Z
ZU
[O(U)]F exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local
)
rewrite with bosonic field φ
〈O(U)〉 = 1Z
ZU,φ,φ†
[O(U)]F exp(−SG(U)− Nf2φ†[(D(U) + m0)
†(D(U) + m0)]−1φ)
≡ 1Z
ZU,φ,φ†
[O(U)]Fe−SPF
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
introduce a (classical) hamiltonian with U as coordinates and SPF as potential andconjugate variables to U (momenta p)
H =Xx,µ
12p2(x , µ) + SPF(U, φ)
algorithm
I U
I p, φ with P(φ) ∝ e−SPF(U,φ), P(p) ∝ e−12
p2
I Hamiltonian evolution
(of course discretisedwith step-size δt)
dU
dt=
dH
dp= p
dp
dt=
−dH
dU= −−dSPF
dU
I U′; metropolis accept/reject (corrects step size errors)
−→ P(U) ∝ e−Seff (U)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
introduce a (classical) hamiltonian with U as coordinates and SPF as potential andconjugate variables to U (momenta p)
H =Xx,µ
12p2(x , µ) + SPF(U, φ)
algorithm
I U
I p, φ with P(φ) ∝ e−SPF(U,φ), P(p) ∝ e−12
p2
I Hamiltonian evolution
(of course discretisedwith step-size δt)
dU
dt=
dH
dp= p
dp
dt=
−dH
dU= −−dSPF
dU
I U′; metropolis accept/reject (corrects step size errors)
−→ P(U) ∝ e−Seff (U)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
introduce a (classical) hamiltonian with U as coordinates and SPF as potential andconjugate variables to U (momenta p)
H =Xx,µ
12p2(x , µ) + SPF(U, φ)
algorithm
I U
I p, φ with P(φ) ∝ e−SPF(U,φ), P(p) ∝ e−12
p2
I Hamiltonian evolution
(of course discretisedwith step-size δt)
dU
dt=
dH
dp= p
dp
dt=
−dH
dU= −−dSPF
dU
I U′; metropolis accept/reject (corrects step size errors)
−→ P(U) ∝ e−Seff (U)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I the expensive step
−dSPF
dU= Nf
2φ†[(D(U) + m0)
†(D(U) + m0)]−1 dD
dU(D(U) + m0)
−1φ)
1. inversion of operator is expensive ∼ V4m−1q
2. dSPFdU→ large for small mq
−→ small δt in discretization of evolution
effort ∝ (δt)−1 ∝ m−nq , n>∼2
total effort ∝ V4 m−nq , n>∼3
(in finite volume, L4, (with Dirichlet B.C.’s) total effort ∝ (L/a)−(4+n) , n>∼3
independent of mq SF simulations −→ Knechtli)
I in general: small quark masses are very difficult to reach
I reach up+down quark masses from heavier onesguided by ChPT → later
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The costs: e.g. for a = 0.08 fm, 1000 configurations Ui
(0.05 fm a factor ∼30)
I Berlin “wall” ——–Lattice 2001, Berlin
disappeared through
Mass preconditioning(“Hasenbusch trick”)[Hasenbusch; Jansen & Hasenbusch; ... ]
multiple time scales[Urbach et al., 2005 ]
Tflops · years
staggered ref [13]
ref [12]
this work
0.00
mPS/mV
10.50
1
0
I more radical: domain decomposition [Luscher ] −→ Giustiprofiting from locality of QCD
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The costs: e.g. for a = 0.08 fm, 1000 configurations Ui
(0.05 fm a factor ∼30)
I Berlin “wall” ——–Lattice 2001, Berlin
disappeared through
Mass preconditioning(“Hasenbusch trick”)[Hasenbusch; Jansen & Hasenbusch; ... ]
multiple time scales[Urbach et al., 2005 ]
Tflops · years
staggered ref [13]
ref [12]
this work
0.00
mPS/mV
10.50
1
0
I more radical: domain decomposition [Luscher ] −→ Giustiprofiting from locality of QCD
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The costs: e.g. for a = 0.08 fm, 1000 configurations Ui
(0.05 fm a factor ∼30)
I Berlin “wall” ——–Lattice 2001, Berlin
disappeared through
Mass preconditioning(“Hasenbusch trick”)[Hasenbusch; Jansen & Hasenbusch; ... ]
multiple time scales[Urbach et al., 2005 ]
Tflops · years
staggered ref [13]
ref [12]
this work
0.00
mPS/mV
10.50
1
0
I more radical: domain decomposition [Luscher ] −→ Giustiprofiting from locality of QCD
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The quenched approximation
〈O(U)〉 = 1Z
ZU
[O(U)]F| z [D(U)+m0]−1
exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local drop it
)
Nf → 0 limit, no sea quarks−→
I quenchedhadron mass spectrumcont. limt[CP-PACS,2002 ]
0
0.5
1
1.5
mh [GeV]
π
K
ρ
K*φ
N
Λ
ΞΣ ∆
Σ∗
Ξ∗
Ω
input: mπ, mρ, mφ
I Amazing success, but qQCD is not QCD. typically 90+% of the answerFactor > 100 work for the remaining 10-x% no better way?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The quenched approximation
〈O(U)〉 = 1Z
ZU
[O(U)]F| z [D(U)+m0]−1
exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local drop it
)
Nf → 0 limit, no sea quarks−→
I quenchedhadron mass spectrumcont. limt[CP-PACS,2002 ]
0
0.5
1
1.5
mh [GeV]
π
K
ρ
K*φ
N
Λ
ΞΣ ∆
Σ∗
Ξ∗
Ω
input: mπ, mρ, mφ
I Amazing success, but qQCD is not QCD. typically 90+% of the answerFactor > 100 work for the remaining 10-x% no better way?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
The quenched approximation
〈O(U)〉 = 1Z
ZU
[O(U)]F| z [D(U)+m0]−1
exp(−SG(U) + Nf Tr ln(D(U) + m0)| z non−local drop it
)
Nf → 0 limit, no sea quarks−→
I quenchedhadron mass spectrumcont. limt[CP-PACS,2002 ]
0
0.5
1
1.5
mh [GeV]
π
K
ρ
K*φ
N
Λ
ΞΣ ∆
Σ∗
Ξ∗
Ω
input: mπ, mρ, mφ
I Amazing success, but qQCD is not QCD. typically 90+% of the answerFactor > 100 work for the remaining 10-x% no better way?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Heavy quarks: The particular challenge for B-physics
multiple scale problemalways difficultfor a numerical treatment
I Take a large lattice as it is possible in the quenched approximation
λπ = 1/mπ ≈ L
λB = 1/mb < a
light quarks are too light→ treat by an extrapolation
b-quark is too heavy
I Need an effective theory for the b-quark: HQET, NRQCD[E. Eichten, 1988; E. Eichten & B. Hill 1990; Caswell & Lepage; Lepage & Thacker ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Heavy quarks: The particular challenge for B-physics
multiple scale problemalways difficultfor a numerical treatment
I Take a large lattice as it is possible in the quenched approximation
λπ = 1/mπ ≈ L
λB = 1/mb < a
light quarks are too light→ treat by an extrapolation
b-quark is too heavy
I Need an effective theory for the b-quark: HQET, NRQCD[E. Eichten, 1988; E. Eichten & B. Hill 1990; Caswell & Lepage; Lepage & Thacker ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Heavy quarks: The particular challenge for B-physics
multiple scale problemalways difficultfor a numerical treatment
I Take a large lattice as it is possible in the quenched approximation
λπ = 1/mπ ≈ L
λB = 1/mb < a
light quarks are too light→ treat by an extrapolation
b-quark is too heavy
I Need an effective theory for the b-quark: HQET, NRQCD[E. Eichten, 1988; E. Eichten & B. Hill 1990; Caswell & Lepage; Lepage & Thacker ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I b-quark istoo heavy:
aMb ≈ 4aMc !
charm: [Sint & Rolf ]
3 definitions ofcharm quark massdiffer by a-effects
charm just doable
a possible alternative(under investiga-tion): Symanzikimprovement toall orders in am[Aoki,Kuramashi ]∼ Fermilab action [El
Khadra,Kronfeld,Mackenzie
]
I for b-quarks a→ 0 can’t be controlled in this way
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
I b-quark istoo heavy:
aMb ≈ 4aMc !
charm: [Sint & Rolf ]
3 definitions ofcharm quark massdiffer by a-effects
charm just doable
a possible alternative(under investiga-tion): Symanzikimprovement toall orders in am[Aoki,Kuramashi ]∼ Fermilab action [El
Khadra,Kronfeld,Mackenzie
]
I for b-quarks a→ 0 can’t be controlled in this way
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Renormalization (with mixing)
1. Bare parameters of QCD, g0,m0,i
→ fundamental parameters, Λ,M0,i ( RGI’s).
prediction of Λ,M0,i with hadron spectrum as input. −→ Knechtli
2. Renormalization of composite operators
e.g. 〈K0|Hweak|K0〉SM ↔ C(µ)〈K0|O∆s=2(µ)|K0〉QCD
↑matching
similar
〈B|Abu0 |0〉 ↔ C(µ)〈B|Abu,stat
0 (µ)|0〉+ O(1/mb) , µ = mb
I matching usually done in the continuum, dimensional regularization,MS-renormal. of coupling and massesstill have to renormalize the bare lattice operators and connect to MS
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Renormalization (with mixing)
3. Weak currents(AR)µ = ZA(g0)Aµ
I special case of composite operators: renormalization is scaleindependent
I can be fixed by χ WI’s [Bochicchio et al. ]I ZA = 1 if the action has exact (lattice) χ symmetry
4. Coefficients in effective (lattice) actions
I e.g. lattice HQET5. Others, in particular hard problems like
OR = Z(µ)
O +
c1(g0)
aO1 +
c2(g0)
a2O2 +
c3(g0)
a3O3
ffI e.g. ∆I = 1/2 problem
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Renormalization (with mixing)
1.A Take a physical observable, expandable in renormalized perturbation theory, e.g.
F (r) =4
3
1
r2
˘αMS(µ) + c1[αMS(µ)]2 + . . .
¯, µ = 1/r
≡4
3
1
r2αqq(µ) physical coupling
compute the physical coupling for large µ; → Λin practice: the Schrodinger functional → Knechtli
1.B Use bare (UV) quantity, expandable in bare (lattice) perturbation theory
e.g. P = 1N〈 trU(p)〉 α ≡ − 1
CFπln( P|z
from MC!
)
then αMS(s0a−1) = α + 0.614α3
+ O(α40) + O(a)
or (see later) αV
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
properties
A) “physical” observable
+ only continuum PT
+ separates continuum limit (an) and renormalization effects (αm)
– needs “window”L−1 Λ µ a−1
+ modification L−1 ≡ µ (use the finite volume); then L a issufficient;large µ by recursion µ0 → 2µ0 → . . .→ 2nµ0
B) UV (cutoff) quantity
+ easy, quick (all you need is a hadron mass)
– mixing of discretization errors (an) and renormalization effects (αm):αMS(s0a
−1)
– difficult (for Nf > 0 impossible) to reach high µ
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
2. Renormalization of composite operatorsExample
O(x) ≡ O∆s=2(x) = s(x)γLµd(x) s(x)γL
µd(x)
assume exact χ symmetry → OR(x) = ZO (µ)O(x)a possible renormalization condition, intermediate scheme
(P(x) = d(x)γ5s(x), P+(x) = s(x)γ5d(x))
〈PR(x1)OR(0)PR(x2)〉 = Z2P(µ)ZO (µ)〈P(x1)O(0)P(x2)〉 e.g. x1 = −x2
= 〈P(x1)O(0)P(x2)〉tree level x21 = x2
2 = µ−2 = x2
〈PR(x)P+R(0)〉 = Z2
P(µ)〈P(x)P(0)〉 = 〈P(x)P(0)〉tree level
? gauge invariant
? for large µ connected to other schemes
OR(µ) = OMS(µ)× [1 + c1αMS + . . .]
? problems of such a ren. cond.– signals in MC– need PT (c1 . . .)– need window: L−1 Λ µ a−1
→ not practical
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Renormalization (with mixing)
alternative 1:
? fix gauge
? P(x) = d(x)γ5s(x)→ d(y1)s(y2) “quark states”
? go to momentum space → MOM-scheme (“RI-MOM”)
– on-shell improvement does not work
– off-shell improvement with non-gaugeinvariant terms
+ rather simple, “universal”
alternative 2:
? renormalization in finite volume, in particular in the SFL a is sufficient;large µ by recursion
+ in the SF: use boundary quark fields instead of P(x)
+ gauge invariant
+ on-shell improvement does work
– need to do PT ourselves (usually 1-loop, sometimes 2)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Renormalization (with mixing)
alternative 1:
? fix gauge
? P(x) = d(x)γ5s(x)→ d(y1)s(y2) “quark states”
? go to momentum space → MOM-scheme (“RI-MOM”)
– on-shell improvement does not work
– off-shell improvement with non-gaugeinvariant terms
+ rather simple, “universal”
alternative 2:
? renormalization in finite volume, in particular in the SFL a is sufficient;large µ by recursion
+ in the SF: use boundary quark fields instead of P(x)
+ gauge invariant
+ on-shell improvement does work
– need to do PT ourselves (usually 1-loop, sometimes 2)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
4. Coefficients in effective (lattice) actions
Coefficients in effective (lattice) actions e.g. lattice HQET
SQCD(. . . ,M) ↔ SHQET(. . . , cσ·B(g0), c2D(g0), . . .)
↑matching
Matching, in principle (M = mb):
OQCD(. . . ,M) = OHQET(. . . , cσ·B(g0), c2D(g0), . . .)
+O(( |p|M )n, ( ΛM )n) + O(am)
This may be done non-perturbatively [Heitger, S. ](again using a finite volume for the matching step)
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Mixing of operators of different dimension
appears e.g. in effective theories such as HQET
OR = Z (µ)
O +
c(g0)
aO1 +
d(g0)
a2O2
perturbation theory is not enough for c , d :
PT: c = c1 g20 + c2g
40 + ...+ cng
2n0 (take d = 0)
I perturbative uncertainty in OR:
δOR ∼g2n+20
a∼ 1
2b0 ln(1/aΛ)n+1 a→∞
no continuum limit
I it is a more general QCD problem: clean computation ofpower corrections needs full non-perturbative treatment of theleading term
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
Mixing of operators of different dimension
appears e.g. in effective theories such as HQET
OR = Z (µ)
O +
c(g0)
aO1 +
d(g0)
a2O2
perturbation theory is not enough for c , d :
PT: c = c1 g20 + c2g
40 + ...+ cng
2n0 (take d = 0)
I perturbative uncertainty in OR:
δOR ∼g2n+20
a∼ 1
2b0 ln(1/aΛ)n+1 a→∞
no continuum limit
I it is a more general QCD problem: clean computation ofpower corrections needs full non-perturbative treatment of theleading term
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
And it is possible: e.g. 1/m corrections in HQET
it is possible with full non-perturbative renormalization
experiment Lattice with amq 1
mB = 5.4 GeV Φ1(L1,M),Φ2(L1,M)
? ?
ΦHQET1 (L2),Φ
HQET2 (L2) ΦHQET
1 (L1),ΦHQET2 (L1)σm(u1)
σkin1 (u1), σ
kin2 (u1)
L2 = 2L1
[ LPHAACollaboration : Heitger & S; Heitger & Wennekers; Heitger, Juttner,S & Wennekers ]
very recent (quenched!): mb, MS-scheme, cont.lim.
mstatb (mb) = 4350(64) MeV ← ±O(Λ2/mb)
mstatb (mb) + m1
b(mb) = 4301(70) MeV ← ±O(Λ3/m2b)
[ LPHAACollaboration : Della Morte, Garron, Papinutto, S. ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ hmc c&b c&b Ren pow corr
And it is possible: e.g. 1/m corrections in HQET
it is possible with full non-perturbative renormalization
experiment Lattice with amq 1
mB = 5.4 GeV Φ1(L1,M),Φ2(L1,M)
? ?
ΦHQET1 (L2),Φ
HQET2 (L2) ΦHQET
1 (L1),ΦHQET2 (L1)σm(u1)
σkin1 (u1), σ
kin2 (u1)
L2 = 2L1
[ LPHAACollaboration : Heitger & S; Heitger & Wennekers; Heitger, Juttner,S & Wennekers ]
very recent (quenched!): mb, MS-scheme, cont.lim.
mstatb (mb) = 4350(64) MeV ← ±O(Λ2/mb)
mstatb (mb) + m1
b(mb) = 4301(70) MeV ← ±O(Λ3/m2b)
[ LPHAACollaboration : Della Morte, Garron, Papinutto, S. ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
Effective field theories for (lattice) QCD
|pµ| Mlow energy, small momentum expansionsderivative expansions e.g.LHQET = ψhD0ψh + 1
mbψhDkDkψh
I M = mb HQET
I M = a−1 Symanzik effective theory
I M = T = Temperature dimensional reduction
I M = 4πFπ and B ×mquark = O(|pµ|2), B = −〈ψψ〉/F 2π ChPT
renormalizable order by order in the expansion → will yield( finite number of counterterms) the asymptotic expansion in |pµ|/Mvery useful (even necessary?)
new developments: combination of ChPT and Symanzik effective theory:
WChPT, SChPT
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
Symanzik effective theory
At energies, momenta far below the cutoff 1/a, lattice QCD is equivalent to acontinuum theory with effective action (on shell)
Seff = SQCD + aS1 + a2S2 + . . .
S1 =
Zd4x c(g0) ψ σµνFµνψ + less relevant
S2 =
Zd4x c ′(g0) trDρFµνDρFµν + ...
⇒ O(a) cutoff effects in on-shell matrix elements can be canceled by adding
a5∑
x
i4cswψσµνFµνψ, csw = 1 + O(g2
0 ),
to the lattice action [Symanzik, Luscher & Weisz, Sheikholeslami & Wohlert,..., LPHAACollaboration ]
I ψ σµνFµνψ is not chirally symmetric (so is the Wilson action)
I other actions exist, which have enough chiral symmetry to enforce c(g0) = 0“automatic O(a) improvement”
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
Symanzik effective theory
At energies, momenta far below the cutoff 1/a, lattice QCD is equivalent to acontinuum theory with effective action (on shell)
Seff = SQCD + aS1 + a2S2 + . . .
S1 =
Zd4x c(g0) ψ σµνFµνψ + less relevant
S2 =
Zd4x c ′(g0) trDρFµνDρFµν + ...
⇒ O(a) cutoff effects in on-shell matrix elements can be canceled by adding
a5∑
x
i4cswψσµνFµνψ, csw = 1 + O(g2
0 ),
to the lattice action [Symanzik, Luscher & Weisz, Sheikholeslami & Wohlert,..., LPHAACollaboration ]
I ψ σµνFµνψ is not chirally symmetric (so is the Wilson action)
I other actions exist, which have enough chiral symmetry to enforce c(g0) = 0“automatic O(a) improvement”
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
Symanzik effective theory
At energies, momenta far below the cutoff 1/a, lattice QCD is equivalent to acontinuum theory with effective action (on shell)
Seff = SQCD + aS1 + a2S2 + . . .
S1 =
Zd4x c(g0) ψ σµνFµνψ + less relevant
S2 =
Zd4x c ′(g0) trDρFµνDρFµν + ...
⇒ O(a) cutoff effects in on-shell matrix elements can be canceled by adding
a5∑
x
i4cswψσµνFµνψ, csw = 1 + O(g2
0 ),
to the lattice action [Symanzik, Luscher & Weisz, Sheikholeslami & Wohlert,..., LPHAACollaboration ]
I ψ σµνFµνψ is not chirally symmetric (so is the Wilson action)
I other actions exist, which have enough chiral symmetry to enforce c(g0) = 0“automatic O(a) improvement”
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
The Zoo of fermion actionsI Wilson: SUL(Nf)× SUR(Nf)× UV (1) broken → SUV (Nf)× UV (1)
O(a)-effectsχ symmetry only in the continuum limit keep m0 > O(aΛ2)mixing of operators of different dimension and of different chirality
I O(a)-improved Wilson: the same but O(a2) effects −→[Sheikholeslami & Wohlert; ...; LPHAA
Collaboration ]I tmQCD: mass term (per doublet) mψγ5τ
3ψ [Frezzotti, Grassi, Sint & Weisz ]simplified O(a) improvementsolves many mixing problemsbreaks parity to PF breaks flavor symmetryat maximum twist: automatic O(a) improvement: [Frezzotti & G.C.Rossi ]
I Domain Wall fermions [Kaplan; Furman & Shamir ]good χ symmetry expensive (factor ≈ 20)
I classically perfect [Hasenfratz & Niedermayer ]good χ symmetry expensive (factor ≈ 10− 50)small a-effects −→
I Ginsparg-Wilson type [Laliena, Hasenfratz, Niedermayer; Neuberger ]exact χ symmetry [Luscher ] very expensive (factor ≈ 10− 100)automatic O(a) improvement
I staggered [Kogut&Susskind ] 4 tastes, but often reduced to 1 by “4-th root trick”(vector) flavor symmetry broken (pions with different masses)
I highly improved; FLIC, ...
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
Chiral perturbation theory including a-effectslow energy effective Lagrangian for QCD: χ Lagrangian
[Weinberg; Gasser & Leutwyler ]
L = L2 + L4 + . . .
L2 =F 2
4Tr
“∂µΣ†∂µΣ
”+
F 2
2Tr
“χ†Σ + χΣ†
”L4 = . . .+ L4 Tr
“∂µΣ†∂µΣ
”Tr
“Σ†χ+ χ†Σ
”+L5 Tr
h∂µΣ†∂µΣ
“Σ†χ+ χ†Σ
”i+L6
hTr
“Σ†χ+ χ†Σ
”i2+ L7
hTr
“Σ†χ− χ†Σ
”i2
+L8 Tr“χ†Σχ†Σ + χΣ†χΣ†
”+ . . .+ L12 Trχχ†
with
M ≡ diag(mu ,md ,ms) Σ ≡ exp
„i2Taπa
F
«χ ≡ 2MB
and a− terms: aψ σµνFµνψ → Tr“∂µΣ†∂µΣ
”Tr
“Σ†ρ+ ρ†Σ
”+ . . .
ρ = a cρ(g0) 1
[Sharpe & Singleton 1998; Sharpe & Shoresh 2002; Rupak & Shoresh 2002 ]
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future Summ Sym WChPT
and a− terms: aψ σµνFµνψ → tr(∂µΣ†∂µΣ
)tr
(Σ†ρ+ ρ†Σ
)+ . . .
ρ = a cρ(g0) 1
Predicts, for mq ΛQCD , a 1/ΛQCD , e.g. (quarks degenerate)
m2π
µ= 1 + γ︸︷︷︸
↑
logµ+ α︸︷︷︸Li
+ α′︸︷︷︸Li
ρ+ ω︸︷︷︸Wi
ρ+ O(µ2, µρ, ρ2)
prediction low energy constants: from fits
µ = Bmq, ρ = acρ(g0)
when mq, a are small enough:
functional formfor extrapolations
mq → mu,md ,
a → 0is known
There are of course extensions to O(a2) very active right now.
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
I some transparencies taken from C. DaviesHEP2005 International Europhysics Conference on High Energy Physics
Lattice QCD 2005Christine Davies
University of Glasgow
Lattice 2005 happening in Dublin
See www.tcd.ie/lat05/index.php
HEPP-EPS 2005
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?I
Take-home message
¥ There has been a revolution in lattice QCD
since 2003
¥ Quenched approximation (ignores sea
quarks) is dead - stop quoting results from it
¥ Lattice QCD now delivering fully
unquenched results: hadron masses that
agree with expt; precise parameters of QCD;
matrix elements relevant to CKM physics.
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IImproved staggered quarks do not solve doubling problem
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
EIGENVALUES
i en alues
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
i ali
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
CHIRALITY
Follana et al (2004-5). . G.P. Lepage, High-Precision LQCD (April 2005). Ð p. 21/43
4 ÔtastesÕ of quark instead of 1 - if all same, Ôdivide by 4Õ :
det(M) → det1/4(M)
ÔTaste-changingÕ interactions mess this up, but vanish as a2
NEW theoretical work encouraging, e.g. eigenvalues
divides into quartets, Index Theorem obeyed etc.
1
4
Follana et al; Durr et al; Adams; Peardon et al; Shamir
(M)
I No Lagrangian formulation known for this “fourth root trick”a discussion in the lattice community:
Q: what are the quark fields?Q: is there a continuum limt? is it QCD?
A: comparison to experiment Q: a good model?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IImproved staggered quarks do not solve doubling problem
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
EIGENVALUES
i en alues
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
i ali
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
2 4 6 8 10 12 14 16
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.2
0.4
0.6
0.8
1
2 4 6 8 10 12 14 16
CHIRALITY
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
FAT7xASQ
ASQTAD
ONE-LINK
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
2 4 6 8 10 12 14 16
EIGENVALUES
Wilson Glue (a ≈
o e Glue (a ≈
o e Glue (a ≈
CHIRALITY
Follana et al (2004-5). . G.P. Lepage, High-Precision LQCD (April 2005). Ð p. 21/43
4 ÔtastesÕ of quark instead of 1 - if all same, Ôdivide by 4Õ :
det(M) → det1/4(M)
ÔTaste-changingÕ interactions mess this up, but vanish as a2
NEW theoretical work encouraging, e.g. eigenvalues
divides into quartets, Index Theorem obeyed etc.
1
4
Follana et al; Durr et al; Adams; Peardon et al; Shamir
(M)
I No Lagrangian formulation known for this “fourth root trick”a discussion in the lattice community:
Q: what are the quark fields?Q: is there a continuum limt? is it QCD?
A: comparison to experiment Q: a good model?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IUpdate on light decay constants
Improved chiral
extraps for
fK
f!= 1.200(4)(+17
−5 )
Vus = 0.2238(+12−32)
CKM 2005 world
average: 0.2262(23)(Blucher)
Use staggered chiral pert. th. to extrap. to physical pion mass (Aubin + Bernard) to take account of disc. errors. NEW 2005 - configswith second ms.
f!, fK
MILC, Aubin et al, hep-lat/0407028; Heller,
this meeting, Bernard LAT05.
!(K → µ") =G2FmKm
2
µ
8#(1−
m2µ
m2K) f 2KV
2
us
Assuming Vud
I fit with 20 free parametersm2
q terms and many terms from StaggeredChPT
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IUpdate on light decay constants
Improved chiral
extraps for
fK
f!= 1.200(4)(+17
−5 )
Vus = 0.2238(+12−32)
CKM 2005 world
average: 0.2262(23)(Blucher)
Use staggered chiral pert. th. to extrap. to physical pion mass (Aubin + Bernard) to take account of disc. errors. NEW 2005 - configswith second ms.
f!, fK
MILC, Aubin et al, hep-lat/0407028; Heller,
this meeting, Bernard LAT05.
!(K → µ") =G2FmKm
2
µ
8#(1−
m2µ
m2K) f 2KV
2
us
Assuming Vud
I fit with 20 free parametersm2
q terms and many terms from StaggeredChPT
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
Concerning StaggeredChPTI Necessary because one staggered field describes 4 quarks, but flavor symmetry is
broken
FIG. 9: Squared masses of charged pions for various tastes on the coarse lattices. We use r1 to set
the scale. Tastes that are degenerate by SO(4) symmetry are fit together.
73
r1 ∼ 1/600MeV−→ splittings of450 MeV
SChPT applicable?with the quoted preci-sion?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
Concerning StaggeredChPTI Necessary because one staggered field describes 4 quarks, but flavor symmetry is
broken
FIG. 9: Squared masses of charged pions for various tastes on the coarse lattices. We use r1 to set
the scale. Tastes that are degenerate by SO(4) symmetry are fit together.
73
r1 ∼ 1/600MeV−→ splittings of450 MeV
SChPT applicable?with the quoted preci-sion?
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IDetermining Parameters of QCD :
nf = 3
nf = 00.2
0.4
0.6
0.8
α(n
f)
V(d/a)
2 4 6 8
d/a(G eV)
Combine 3-loop
pert. th. for 28
different
quantities (mainly
Wilson loops)
with numerical
calc. on lattice
!s
Results at 3 values of a allows estimates of 4-loop
terms. d/a is BLM scale - differs for each W, so see
running of .
Wlatt =3
!n=1
cn"n
V(d/a)
Quenched
Unquenched
!s HPQCD, Mason et al, hep-lat/0503005
I log W12 = −5.551αV (3/a)× [1− 0.86α+ 1.72α2−5(2)| z ↑
α3−1(2)| z ↑
α4 + . . .]
α = 0.25 . . . 0.4 fitted
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IDetermining Parameters of QCD :
nf = 3
nf = 00.2
0.4
0.6
0.8
α(n
f)
V(d/a)
2 4 6 8
d/a(G eV)
Combine 3-loop
pert. th. for 28
different
quantities (mainly
Wilson loops)
with numerical
calc. on lattice
!s
Results at 3 values of a allows estimates of 4-loop
terms. d/a is BLM scale - differs for each W, so see
running of .
Wlatt =3
!n=1
cn"n
V(d/a)
Quenched
Unquenched
!s HPQCD, Mason et al, hep-lat/0503005
I log W12 = −5.551αV (3/a)× [1− 0.86α+ 1.72α2−5(2)| z ↑
α3−1(2)| z ↑
α4 + . . .]
α = 0.25 . . . 0.4 fitted
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IDetermining parameters of QCD: mq
Use pert. th. to convert
lattice bare mass to
New b/c masses (GeV):
mb(mb) = 4.4(3)
mc(mc) = 1.10(13)
Light quark masses now
have 2-loop matching! 1
10
100
1000
10000
Quark masses (MeV)
u
d
s
c
b
2005 lattice QCD2004 PDG MS
HPQCD, Gray et al, hep-lat/0507013,
Nobes et al, LAT05;
N.B. 3-loop nf=2 result mb = 4.21(7) from
Direnzo and Scorzato, hep-lat/0408015
ms = 86(5); mu = 1.9(3); md = 4.5(4) in MeV at 2 GeV
HPQCD/MILC, Aubin et al, hep-lat/0405022; HPQCD, Mason, LAT05 updates this with 2-loop results
I what is the influence of the taste violations to these numbers?
I there is progress in NP renormalization −→ Knechtli
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IDetermining parameters of QCD: mq
Use pert. th. to convert
lattice bare mass to
New b/c masses (GeV):
mb(mb) = 4.4(3)
mc(mc) = 1.10(13)
Light quark masses now
have 2-loop matching! 1
10
100
1000
10000
Quark masses (MeV)
u
d
s
c
b
2005 lattice QCD2004 PDG MS
HPQCD, Gray et al, hep-lat/0507013,
Nobes et al, LAT05;
N.B. 3-loop nf=2 result mb = 4.21(7) from
Direnzo and Scorzato, hep-lat/0408015
ms = 86(5); mu = 1.9(3); md = 4.5(4) in MeV at 2 GeV
HPQCD/MILC, Aubin et al, hep-lat/0405022; HPQCD, Mason, LAT05 updates this with 2-loop results
I what is the influence of the taste violations to these numbers?
I there is progress in NP renormalization −→ Knechtli
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
A recent revolution?
IDetermining parameters of QCD: mq
Use pert. th. to convert
lattice bare mass to
New b/c masses (GeV):
mb(mb) = 4.4(3)
mc(mc) = 1.10(13)
Light quark masses now
have 2-loop matching! 1
10
100
1000
10000
Quark masses (MeV)
u
d
s
c
b
2005 lattice QCD2004 PDG MS
HPQCD, Gray et al, hep-lat/0507013,
Nobes et al, LAT05;
N.B. 3-loop nf=2 result mb = 4.21(7) from
Direnzo and Scorzato, hep-lat/0408015
ms = 86(5); mu = 1.9(3); md = 4.5(4) in MeV at 2 GeV
HPQCD/MILC, Aubin et al, hep-lat/0405022; HPQCD, Mason, LAT05 updates this with 2-loop results
I what is the influence of the taste violations to these numbers?
I there is progress in NP renormalization −→ Knechtli
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The future
is bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The futureis bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The futureis bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The futureis bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The futureis bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice
Why Basics Chall EFTs Rev Future
The futureis bright
I already now: large impact of LQCD, in particular onCKM physics −→ Lubiczdetermination of fundamental parameters −→ Knechtli
I continuous development of theoretical tools, conceptionally newideasLattice PT, NP Symanzik improvement, NP renormalization,physics from finite size effects, exact chiral symmetry, ChPTincluding lattice effects ...
I progress in algorithms small quark masses −→ Giusti– often driven more by intuition than by theoretical understanding
a very complex dynamical system −→ a challenge– work is not sufficiently appreciated
I driven also by advances in computer technology
I there is lots to do
Rainer Sommer Introduction to QCD on the lattice